Attribution vs Incrementality: The Playoffs

Over many years, the art of ‘attribution modelling’ for predominately digital media has focused on rewarding the correct amount of credit for each channel based on each channel’s ability to play a role within a conversion or a sale. The metaphor that is often used to describe this process is one of a football team and awarding each player credit for the teams’ performance rather than simply the goal scorer. The analogy is that instead of the goal scorer getting all the glory, the goalkeeper is also given credit for defending against an opposition attack, the defender for booting the ball forward and the right winger for sending in a great cross.

This debate, is often about rewarding channels at the top of the funnel who are generating demand for a product or service as well as those at the bottom of the funnel just before sale. And to put it another way - not simply giving Google PPC all the cash because they were the ‘last click’ before the goal conversion but redistributing the benefits through the marketing eco system to channels that generate the demand as well such as display and video media.

Research has focused on testing and debating the merits of first click attribution vs last click or propensity attribution based on non-converting vs converting user journey’s and clients have waded through a plethora of vested interests and complexity around ‘trackable’ channels vs ‘untrackable’ channels, online vs offline decision making or upper funnel tactics with long lead times prior to sale vs conversion actions which demonstrate effects in the short term.

The benefits of the attribution approach is that it helps avoid the common marketing dilemma of over focusing on short term efficiency gains at the expense of long term effectiveness – or optimising yourself into a corner. On the flip side, digital trackable channels are over rewarded at the expense of harder to track but potential higher reach media such as TV and OOH.

I have noted, however, that over the last few years the debate has shifted again and moved beyond attribution as an ecosystem to trying to understand the incrementality of each particular channel within a marketing plan. To over torture the football analogy; the issue with attribution studies is that whilst many more players (or channels) on a pitch are more likely to win the match and they can all provide evidence for supporting the goal scorer to success, attribution modelling tends to say all channels have a role to play to a greater or lesser extent and not which channels can be dropped with minimal effects. For all those, who have sat through presentations from digital media vendors who have claimed that they contributed to 40% of sales and thought ‘what really? So if I didn’t spend with you my sales would go down by 40%?’; tested it and found no differences’ then a methodology is required to understand which team players or channels are truly the critical ones to win the game and therefore the ones that are worth spending the premium transfer fees on.

Methods to try and uncover these incremental effects hark back to the old days of traditional direct response advertising: test and control groups for each channel combination. Thus a plan of Facebook retargeting is tested vs PPC only with different test and control groups assigned and the outcomes assessed. One typical way of doing this is via the PSA ad test or placing a charity ad instead of the advertiser ad and tested the differences.

Will this method of incrementality become the only route to playing the (digital) marketing game? Perhaps. But the increasing focus on micro testing each channel budget shift can seriously slow down marketing decision making in the real time advertising world. As in all matters of marketing measurement, a balanced approach is required to understand the direction of travel rather than each marketing methodology effectively trumping the one before.

Kate Cox recently gave her top tips from the IPA Effectiveness Awards at a WARC webinar. Entries for the IPA Effectiveness Awards close on 20th April.